Officer Carol Ogrinc

Notes from CALEA Conference

Orlando, Florida

Survey should be agency specific

Objectives must be clear

Must be defined

Have to know what to expect to find

Ask questions in a way so everyone knows what you are asking

Ask appropriate questions

Fulfill CALEA Standards and mandates

A 20% response is considered good

Refine your plan to set priorities for the next three years

Most common

Cost effective. One person can design and administer

Sampling frame (victims & witnesses)

Can get list of names from Tax Appraisers office

Single-family homes

Duplexes – apartments, businesses, mobile homes, etc.

Instead of asking their income, ask a proxy question;

What is the value of your home?

What kind of vehicle do you own/drive?

 

Able to obtain large sample

Are great sources of data

Turn around time is limited

Can purchase telephone lists

Many people use cell phones

-2-

Drawback – if you rely solely on a phone book, you will miss younger people

Will get a 5 – 15% response

Responses are quick

Based on a convenience sample

Drawback – Must have access to a computer (may miss elderly group)

Sotfware available – Microsoft Frontpage

Should have some type of "giveaway" in exchange for their time

Focus Groups

Can ask what important issues are in their neighborhood

What is important to them?

Traffic problems?

 

To measure agency performance

Measure officer performance

Solicit recommendations for improvement

Identify citizen concerns

It is important to listen to people. They feel as though they have input in police/community.

 

Your goal is to get meaningful data

Must include control variables – demographic variables

**Demographic variables (important)

Gender, race, income and education – Is sample reflective of the true population?

-3-

Is data able to be generalized?

Agency performance – be careful how you ask question

Everyone must be able to answer a question with a category that pertains to them.

Mutually exclusive – can only give one answer to a specific question

Try to avoid "check all that apply" – Is difficult to analyze

Likert Scale – make sure you have a "neutral" category in the middle

Good | | | | | Bad

quite slightly neither slightly quite

or

(almost never) 1 2 (sometimes) 3 4 5 (most of the time)

5 point Likert – should give you a normal distribution (graph) scale

(3) no opinion/neutral

Make sure you collect a large enough sample representative of your population

Ideal sample size – at least 384

By: Don A. Dillman

By: Priscilla Salant & Don A. Dillman

By: Robert M. Groves, Don A. Dillman & John L. Eltinge

BACK TO CONNPAC HOME PAGE